Surveying with GPS, total station and

terresterial laser scaner: a

comparative study

Solomon Dargie Chekole

Master of Science Thesis in Geodesy No. 3131

TRITA-GIT EX 14-001

School of Architecture and the Built Environment

Royal Institute of Technology (KTH)

Stockholm, Sweden

May 2014

-2-

Abstract

Today, advanced GPS receivers are improving the accuracy of positioning information, but in

critical locations such as urban areas, the satellite availability is limited above all due to the

signal blocking problem, which degrade the required accuracy. For this reason, different methods

of measurement should be used.

The objective of this thesis is to evaluate and compare precision, accuracy and time expenditure

of total station (TS), Global Positioning System (GPS) and terrestrial laser scaner (TLS).

Comparing precision, accuracy and the required time of these three measurements will improve

the knowledge about how much precision and accuracy can be achieved and at what time

expense. To investigate this task, a reference network consisted of 14 control points has been

measured five times with Leica 1201 TS and served as a reference value for comparison with

RTK and TLS measurements. The reference network points were also measured five times with

the GPS RTK method so as to compare accuracy, precision and time expenditure with that of TS.

In addition, in order to compare the accuracy, precision and time expense of total station and

TLS, the North Eastern façade of the L building at KTH campus in Stockholm, Sweden has been

scaned five times with HDS 2500 scaner on six target points. These six target points were also

measured five times with TS. Then comparison made to evaluate the quality of the coordinates of

the target points determined with both measurements. The data were processed in Cyclone, Geo

Professional School and Leica geo office software.

According to the result obtained, the reference network points measured with TS were

determined with 1 mm precision for both horizontal and vertical coordinates. When using RTK

method on the same reference network points, 9 mm in horizontal and 1.5 cm accuracy in

vertical coordinates has been achieved. The RTK measurements, which were measured five

times, determined with a maximum standard deviation of 8 mm (point I) and 1.5 cm (point A) for

horizontal and vertical coordinates respectively. The precision of the remaining control points is

below these levels.

The coordinates of the six target points measured with TS on the L building façade were

determined with a standard deviation of 8 mm for horizontal and 4 mm for vertical coordinates.

When using TLS for the same target points, 2mm accuracy has been achieved for both horizontal

and vertical coordinates. The TLS measurements, which were measured five times, determined

with a maximum standard deviation of 1.6 cm (point WM3) and 1.2 cm (point BW11) for

horizontal and vertical coordinates respectively. The precision of the remaining control points is

below these levels.

With regard to time expenditure, it is proved that total station consumed more time than the other

two methods (RTK and TLS). TS consumed 82 min more time than RTK but, almost similar

time has been consumed by TS and TLS (38 min for TS and 32 min for TLS).

-3-

Acknowledgement

This thesis would not be possible without the help of some individuals. First and for most I

would like to express my deepest gratitude to my supervisors, Dr. Milan Horemuz and Erick

Asenjo for their help, valuable advice, continuing support, endless patience and guidance

throughout my thesis. I could learn more than I expected and so I am very lucky to be Milan’s

student. I would also like to thank my examiner Professor Lars Sjöberg to review my thesis.

I have special thanks to my colleagues who have helped me throughout my thesis work. I

appreciate their help during the field measurement since working with laser scaner and total

station instrument was very difficult to do alone. They could resist the bad weather condition,

snow, which is very harsh for them. I couldn’t pass without acknowledging my classmate, Maria,

who has contributed a lot to my success.

I am also grateful to Dr. Huaan Fan for his support, encouragement and follow-ups throughout

my study. His overall support encouraged me to complete my study properly and on time.

I greatly acknowledge KTH in collaboration with Institute of Land Administration, Bahir Dar

University for providing me with this scholarship to study at KTH.

-4-

Table of Contents

ABSTRACT .............................................................................................................................................................. 1

ACKNOWLEDGEMENT ............................................................................................................................................ 3

1 INTRODUCTION ........................................................................................................................................... 10

1.1 BACKGROUND OF THE STUDY ............................................................................................................................. 10

1.2 PROBLEM STATEMENT ...................................................................................................................................... 11

1.3 OBJECTIVE OF THE RESEARCH ............................................................................................................................. 12

1.4 SIGNIFICANCE OF THE STUDY .............................................................................................................................. 12

1.5 SCOPE AND LIMITATION OF THE STUDY ................................................................................................................. 12

1.6 THESIS OUTLINE ............................................................................................................................................... 12

2 LITRATURE REVIEW ...................................................................................................................................... 13

3 OVERVIEW OF SURVEYING METHODS .......................................................................................................... 15

3.1 LASER SCANING OVERVIEW ....................................................................................................................................... 15

3.1.1 Registration and Geo-referencing .......................................................................................................... 16

3.2 OVERVIEW OF TOTAL STATION ........................................................................................................................... 17

3.2.1 Measurement accuracy .......................................................................................................................... 18

3.2.2 Measurement Errors............................................................................................................................... 19

3.2.3 Mode of distance measurement ............................................................................................................ 20

3.3 OVERVIEW OF GPS .......................................................................................................................................... 21

3.3.1 Real Time Kinematics (RTK) .................................................................................................................... 22

3.3.2 Comparison of Total Station and GPS .................................................................................................... 22

3.3.3 Comparison of Total Station and Laser Scaner ....................................................................................... 23

3.4 ERROR ANALYSIS .............................................................................................................................................. 24

3.4.1 Measurement Errors............................................................................................................................... 24

3.4.2 Accuracy ................................................................................................................................................. 25

3.4.3 Precision ................................................................................................................................................. 26

3.4.4 Checking accuracy .................................................................................................................................. 26

3.4.5 Quality Control ....................................................................................................................................... 26

4 METHODOLOGIES ........................................................................................................................................ 27

4.1 ESTABLISHING REFERENCE NETWORK .................................................................................................................. 27

4.2 EVALUATION OF ACCURACY AND PRECISION .......................................................................................................... 27

4.2.1 Choosing suitable control points for the network and detail survey ...................................................... 28

4.2.2 Setting up targets for laser scaning ....................................................................................................... 28

4.2.3 Detail survey ........................................................................................................................................... 28

5 RESULTS AND DISCUSSION ........................................................................................................................... 33

5.1 GPS BASELINE PROCESSING .............................................................................................................................. 33

5.2 ADJUSTMENT .................................................................................................................................................. 33

5.2.1 Rounds of Measurement ........................................................................................................................ 33

5.2.2 A priori standard deviation ..................................................................................................................... 33

5.2.3 Horizontal Network Adjustment ............................................................................................................. 34

5.2.4 Adjusted coordinates and their standard deviations ............................................................................. 35

5.2.5 Vertical Network adjustment ................................................................................................................. 37

-5-

5.3 DETERMINATION OF PRECISION AND ACCURACY OF RTK .......................................................................................... 38

5.4 REGISTRATION AND GEO REFERENCING ................................................................................................................ 42

5.5 COMPARISON OF LASER SCANER AND TOTAL STATION RESULT .................................................................................. 43

5.6 COMPARISON OF TIME EXPENDITURE .................................................................................................................. 48

5.6.1 Total Station versus GPS ......................................................................................................................... 48

5.6.2 Laser scaning and total station .............................................................................................................. 49

6 CONCLUSION AND RECOMMENDATION ...................................................................................................... 50

6.1 CONCLUSION .................................................................................................................................................. 50

6.2 RECOMMENDATIONS ........................................................................................................................................ 51

REFERENCES ......................................................................................................................................................... 52

-6-

List of Figures

FIG.3. 1: PICTURE OFHDS 2500 ........................................................................................................... 15

FIG.3. 2: LEICA 1201 TOTAL STATION ................................................................................................... 17

FIG.3. 3: COLLIMATION ERRORS ............................................................................................................ 20

FIG.3. 4: GPS RECEIVER ...................................................................................................................... 22

Fig.4. 1: Reference Control point……..………………………………………………………….30

Fig.5. 1: A priori standard deviation………………………………………………………………………………………….34

FIG.5. 2: GRAPHICAL VIEW OF THE REFERENCE NETWORK ........................................................................... 37

FIG.5. 3: GRAPHICAL VIEW OF HEIGHT ADJUSTMENT ................................................................................. 37

FIG.5. 4: GEO-REFERENCED POINTS....................................................................................................... 42

FIG.5. 5: REGISTERED AND GEO-REFERENCED SCANWORLDS ...................................................................... 43

-7-

List of Tables

TABLE3. 1: CLASSIFICATIONS AND ACCURACY OF LASER SCANNERS BASED ON MEASUREMENT PRINCIPLE .............. 16

TABLE3. 2: ANGLE MEASUREMENT ACCURACY .......................................................................................... 19

TABLE3. 3: ANGLE ERRORS AND THEIR ADJUSTMENT. ................................................................................ 19

TABLE3. 4: RANGE LIMIT BASED ON ATMOSPHERIC CONDITION .................................................................... 20

TABLE3. 5: ACCURACY OF MEASUREMENTS TO STANDARD PRISM ................................................................. 21

TABLE3. 6: DISTANCE ACCURACY IN RL MODE .......................................................................................... 21

TABLE3. 7: COMPARISON OF IR AND RL MODE ........................................................................................ 21

TABLE3. 8: COMPARISON OF GPS AND TOTAL STATION ............................................................................. 23

TABLE3. 9: COMPARISON OF LASER SCANER AND TOTAL STATION ................................................................. 23

TABLE3. 10: CONCLUSION OF PROS AND CONS OF TOTAL STATION GPS AND LASER SCANNER ........................... 24

Table5. 1: Computed coordinates (m)……………………………………………………………33

TABLE 5. 2: SOME OF ADJUSTED VALUES AND THEIR STANDARD DEVIATIONS ................................................... 35

TABLE 5. 3: ADJUSTED COORDINATES OF THE REFERENCE NETWORK ............................................... 36

TABLE 5. 4: RTK MEASUREMENT, ITS RMS AND STANDARD DEVIATIONS ....................................................... 38

TABLE 5. 5: COMPARISON OF STANDARD DEVIATIONS BETWEEN TS AND RTK ................................................ 40

TABLE 5. 6: CONFIDENCE INTERVAL LIMITS AND COORDINATES DIFFERENCE BETWEEN TPS AND RTK .................. 41

TABLE 5. 7: STANDARD DEVIATIONS OF THE REGISTERED TLS ...................................................................... 44

TABLE 5. 8: STANDARD DEVIATIONS OF TS MEASUREMENTS ....................................................................... 45

TABLE 5. 9: STANDARD DEVIATION AND RMS OF TLS MEASUREMENT ............................................ 46

TABLE 5. 10: CONFIDENCE INTERVAL LIMIT FOR THE DIFFERENCE BETWEEN TPS AND TLS ................................. 47

TABLE 5. 11: TIME EXPENDITURE FOR TS AND RTK MEASUREMENTS FOR THE REFERENCE NETWORK .................. 49

TABLE 5. 12: TIME EXPENDITURE FOR TS AND LASER SCANER ON THE FAÇADE ................................................ 49

-8-

Acronyms

3D Three Dimension

BW Black and White target

ATR Automatic Target Recognition

CORS Continuously Operating Reference Station

EDM Electronic Distance Measurement

GNSS Global Navigation Satellite System

GPS Global Positioning System

HDS High Definition Scanner

IR Infrared Reflector

P Point

QC Quality Control

RL Reflector-less

RMS Root Mean Square

RT Red target

RTK Real Time Kinematics

SWEPOS Sweden Positioning System

SWEREF 99 Swedish Reference Frame 99

TLS Terrestrial Laser Scanning

TPS Terrestrial Positioning System

UTM Universal Transverse Mercator

WM Window Mirror corner

-9-

Glossary

Accuracy: refers to how closely a measurement or observation comes to

measure a true or established value

Adjustment: the process of correcting errors made during the measurement.

Control network: is a reference that can be served as a reference value for RTK

and TLS measurements in order to evaluate the accuracy.

GPS: Global positioning system is also a surveying instrument that determines

coordinates of a point relative to WGS 84. Its height reference is the ellipsoid.

Precision: refers to how closely repeated measurements or observations come to

duplicate the measured or observed values.

ScanWorld: is a term used in Cyclone software to refer a scaned scene from one

position of the scaner.

Terrestrial Laser Scaning: can be defined as use of a laser to collect

dimensional data of objects in the form of a point cloud.

Time expenditure: defined as time consumed to perform the required task.

Total station: is a surveying instrument that determines coordinate of a point

indirectly from measured angles and distances. Its height reference is the geoid.

3D quality: is a measure of accuracy, in which, it is calculated using standard

deviations of the 3D coordinates with Eq. (4.3).

Effective time: is the time needed to measure the required tasks without

considering the delayed time (time consumed for changing battery, transporting

instruments, etc.)

-10-

1 INTRODUCTION

1.1 Background of the study

The research deals with evaluation and comparison of precision, accuracy and time expenditure

of three surveying methods. These methods are total station (TS), Global positioning system

(GPS), and terrestrial laser scaner (TLS).

Surveying has been an essential element in the development of the human environment for so

many centuries. It is an imperative requirement in the planning and execution of nearly every

form of construction. Surveying was essential at the dawn of history, and some of the most

significant scientific discoveries could never have been implemented, were it not for the

contribution of surveying. Its principal modern uses are in the fields of transportation,

construction building, apportionment of land, and detail mapping1.

In surveying, specifically in the area of engineering projects, more sophisticated instruments are

employed (total station, laser scaner and GPS) to improve the efficiency and accuracy. Individual

surveying techniques has been commonly used in the history of surveying area to collect data

from field measurements for various applications with different accuracy capabilities and

requirements. The significant development of surveying techniques enabled surveying

professionals to evaluate precision and accuracy of different surveying techniques. As a result of

this evaluation, many advantages has been gained; basically such as improving the efficiency and

accuracy of the results. The accuracy of surveying measurements can be improved almost

indefinitely with increased cost (time, effort and money).

Today, the role of surveying got much attention to be used in many applications with better

accuracy. The term accuracy is common in many applications to express the quality of

observations, measurements or/and calculations.

The required accuracy depends on the needed deliverable output. Applications such as general

navigation tasks on the sea, research in oceanography, position and velocity in small scale

geophysical exploration are required low accuracy, applications such as hydrography, calibration

of transponder system, precise navigation and seismic survey, precise navigation in coastal

waters etc. are grouped as medium accuracy requirements and applications which require high

accuracy are; precise hydrographic surveying, support of coastal engineering marine,

geodynamics, precise continuous height control, engineering construction projects (Sjöberg,

2012).

Accuracy and precision for those in the surveying profession (as well as other technical and

scientific fields) are defined in different way. Accuracy refers to how closely a measurement or

observation comes to measure a true or established value, since measurements and observations

1 http://www.britannica.com/EBchecked/topic/575433/surveying [Accessed 09 February 2013]

-11-

are always subject to errors. Precision refers to how closely repeated measurements or

observations come to duplicate the measured or observed values.

Accuracy of surveying techniques using instruments such as GPS, TS and TLS are dependent on

a number of parameters that limit their measurement quality. For instance: multipath, the inherent

satellite signal accuracy, signal transmission delay, receiver hardware and software limitations,

satellite signal obstruction are some of the problems associated with GPS measurement. On the

other hand, limitations stemming from total station are; computed coordinates are in local or

target coordinate system, the reference surface for measuring height is geoid. Because of earth’s

curvature, the accuracy of TS measurement can also be affected by distance limit (the accuracy

will decrease when increasing the distance). Finally, accuracy of laser scaner depends on the

angle of sight and distance from the object to be scaned i.e. scaning perpendicular to the object is

more accurate than slightly inclined scaning. Even if laser scaner can capture thousands of points

per second, all of these points cannot be handled easily to manipulate and store.

Therefore, each method has its own advantages and disadvantages. In addition to the above

differences, the methods have also different time consumption to do the required tasks. Thus, the

scope of the research is to evaluate and compare accuracy, precision and time expenditure of the

above three methods.

1.2 Problem statement

Surveying is the technique and science of accurately determining three-dimensional position of

points and the distances and angles between them. Various surveying methods (GPS, laser

scaner, total station, etc.) are in use. In this research only these instruments have been used.

The latest geodetic GPS receivers are improving the accuracy of positioning information, but in

critical locations such as urban areas, the satellite availability is difficult due to the signal

blocking problem, multipath etc. which degrade the required accuracy. 3D laser scaners generate

up to thousands of points per second, however, handling and manipulating the huge amount of

point data is a major problem. To avoid these problems, it is very important to reduce the amount

of acquired point data. As a result of this reduction of data, accuracy of the final result will be

altered. Total station can measure a single point coordinate precisely, but the computed

coordinates are in local or target coordinate system, which needs datum transformation. The

accuracy is affected with angle and distance of sight, weather condition, etc.

Considering those limitations, the research will evaluate and compare accuracy, precision and

time expenditure of these three surveying methods (total station, GPS and laser scaner).

-12-

1.3 Objective of the research

The general objective of this research is intended to evaluate and compare the accuracy, precision

and cost (time expenditure) of three methods, i.e. GPS, total station and laser scaning.

Specifically the research intends to:

Determine and evaluate precision of the reference network which can be served as a

reference value for comparison with RTK and TLS

Determine and evaluate accuracy and precision of GPS RTK and TLS methods

Determine the cost (time expenditure) of the three methods

Compare results of the methods based on RMS and standard deviation analysis

Forward possible recommendations that can improve the precision and accuracy of the

three measurement methods

1.4 Significance of the study

This research can be used as a spring board for further studies for those who are

interested in the area. On the other hand, the study can help users to choose appropriate

methods for a given task. Moreover, since coordinates of the reference points are

determined with high precision, it can be serves as a reference values for other users.

1.5 Scope and limitation of the study

The scope of this study is limited within evaluating and comparing the accuracy,

precision and time expenditure of three surveying methods. Determining and evaluating

the accuracy of the measurement need quite stable weather condition and carefulness.

During this work there have been a lot of limitations especially related with whether

condition (cold, snow and wind). Due to this problem, the study couldn’t complete

according to the time frame work.

1.6 Thesis outline

Chapter one introduces the overall background, problem and objective of the thesis.

Chapter two starts with literature review, which describes the overview and fundamentals

of GPS, total station and laser scaner. It also presents other’s related work. Chapter three

introduces methodology of the thesis and procedures. Chapters four presents the result

and discuss the result in detail. Chapter five gives conclusion and recommendation that

can improve the result.

-13-

2 LITRATURE REVIEW

This section describes some of what others have done in related work in order to give brief idea

about the overall concept of precision, accuracy and time expenditure of total station, GPS and

TLS.

According to the work by Ehsani et al, (2004), a 50 ha area was surveyed with RTK-GPS. The

base station and four reference points were established over the highest point in the survey area

Corrected GPS signals are transmitted in real time from a base receiver at a known location to

one or more rover receivers. Results from RTK GPS method, a horizontal coordinate accuracy of

1 cm has been achieved by compensating for atmospheric delay, orbital errors and other variables

in GPS geometry. Comparing this thesis with the above work, 8 mm horizontal coordinate

accuracy achieved using the same method (RTK).

According to Lin, (2004), accuracy test was made between GPS RTK and total station. The

results showed that a positional accuracy of 14 mm has been achieved using GPS RTK while

using total station it was possible to determine 16 mm positional accuracy.

Any blockage from natural or man-made obstacles such as trees and buildings can make use of

RTK method limited or impossible. In such cases, total stations are used. Borgelt et al, (1996)

compared the accuracy of RTK with total station on the free area and they reported a standard

deviation of 12 cm in a vertical position with RTK. But in the case of total station, better results

(below 5 mm) have been achieved.

Pflipsen, (2006), has tested accuracy and time expenditure of total station versus laser scaner on a

pile of sand for comparison purpose. The pile was surveyed twice: once with a laser scaner

(Leica HDS 2500) and once with a total station (Leica TS1200), and he processed the data in

Cyclone and Geo software respectively. His result showed that almost similar horizontal and

vertical coordinate accuracy have been achieved below 9 mm in both methods. The time

consumed for the measurements was a little bit more (7 minute) for the total station.

According to the studies conducted by Jonsson, et al (2003), RTK measurement was applied to

test accuracy of different GPS instruments (Leica, Topcon and Trimble). A network of nine

control points was established using total station. Then, the authors performed RTK

measurement on the same network and compared results with different instrument. Results

obtained from RTK measurement have shown a horizontal and vertical accuracy of 10 mm and 2

cm respectively. When comparing this result with the result of the thesis, better accuracy was

achieved in both horizontal and vertical coordinates.

In order to check the compatibility of the RTK method with that of total station method, Ahmed,

(2012) tested RTK and total station measurements on an existing network. The objective of the

test was to assess the RTK achievable accuracy, to check the repeatability of the results under

-14-

different satellite configurations and to evaluate RTK performance in urban area. In the test,

accuracy and repeatability assessment of the RTK was carried out by comparing the coordinates

of points with that of independently precisely determined using a total station. According to the

result, the difference between the coordinates of total station and RTK was 2 cm for the

horizontal and 3 cm for the vertical coordinates. In comparison with the results of this thesis, the

coordinate difference between total station and RTK (coordinates of RTK- coordinates of TS)

was 1.8 cm for both horizontal and vertical coordinates.

In another study by Fregonese, et al, (2007), the objective of the study was to access the

feasibility of monitoring deformations of large concrete dams using terrestrial laser scaning. For

this purpose a test field has been established on the specific dam. First the author established a

geodetic network as a reference by Leica TS, and then, using a number of targets on the dam,

measurements were taken with a total station and a laser scaner. The reference network was

determined with 2 mm horizontal and 3 mm vertical coordinate precision. Targets, mounted on

the dam, were measured precisely with a total station, and 3 mm for the horizontal and 4.5 mm

for the vertical coordinate accuracy (RMS) has been achieved. On the other hand, using a laser

scaner (HDS 300), 4 mm for the horizontal and 8 mm for the vertical coordinate accuracy (RMS)

was achieved.

-15-

3 OVERVIEW OF SURVEYING METHODS

3.1 Laser Scaning Overview

Laser scaning has been conceived as a method to directly and accurately capture object surfaces.

According to Fazlay, (2003), although 30 years old, the commercial market for laser scaning has

only developed significantly after 1996. Laser scaning is a method where a surface is sampled or

scaned using laser technology. It collects data on the object’s shape and possibly its appearance.

The collected data can then be used to construct digital, two-dimensional drawings or three-

dimensional models useful for a wide variety of applications. The advantage of laser scaning is

the fact that it can record huge numbers of points with high accuracy in a relatively short period

of time. It is like taking a photograph with depth information. Laser scaners are line-of-sight

instruments, so to ensure complete coverage of a structure multiple scan positions are required

(Quintero et al, 2008).

Fig.3.1: picture ofHDS 2500

In this thesis Leica HDS 2500 scaner (Fig.3.1: picture ofHDS 2500) which has a maximum 40° x

40° field of view was used. With a single point range accuracy of +/- 4 mm, angular accuracies

of +/- 60 micro-radians, and a beam spot size of only 6 mm from 0-50 m range, the HDS 2500

delivers survey grade accuracy while providing a versatile platform for data capture. Its 360° x

195° pan and tilt mount and dual internal rotating mirrors enable it to be deployed in virtually

any orientation. The combination of high accuracy and field versatility makes the HDS 2500

ideal for fixed or raised installation when leveled tripod mounting is not practical, or for

applications with less stringent field of view requirements2.

Classification of laser scaners (Table 3.1) based on technical specification and measurement

principle:

– scaning speed, sampling rate of laser measurement system

– field of view (camera view, profiling, imaging)

– spatial resolution, i.e. number of points scaned in field of view

– accuracies of range measurement system and deflection system

2 http://hds.leicageosystems.com/en/5940.htm [Retrieved on March 20, 2013]

-16-

Table 3.1: Classifications and accuracy of laser scaners based on measurement principle

Measurement

technology

Range [m] Accuracy [mm] Manufacturers

Time of flight < 100 < 10 Leica, Mensi, Optech,

Riegl, Callidus

< 1000 < 20 Optech, Riegl

Phase measurement < 100 < 10 IQSun, Leica, VisImage,

Zoller+Fröhlich

Optical triangulation < 5 < 1 Mensi, Minolta

3.1.1 Registration and Geo-referencing

Registration is the process of integrating the ScanWorlds into a single coordinate system. Here

the term ScanWorld is used in Cyclone software to refer to a scaned scene from one station

setup. The scaned scene is a collection of 3D points which can be called as point clouds. The

registration is derived by using a system of constraints, which are pairs of equivalent or

overlapping objects that exist in two ScanWorlds. The registration process computes the optimal

overall alignment transformations for each component (Easting, Northing and height) of

ScanWorld in the registration, such that the constraints are matched as closely as possible. Combining several datasets into a global consistent model is usually performed using

registration. The key idea is to identify corresponding points between the scaned scenes and find

a transformation that minimizes the distance between corresponding points. Registration of point

clouds in the same coordinate system is the most important step in the processing of terrestrial

laser scaner measurements. In order to perform the registration, ScanWorlds have to be

overlapped at least 30% each other.

Data points in a captured dataset from any acquisition system may be associated with specific

reference coordinate system on the earth’s surface. This leads to the term geo-referencing, which

can be defined as “the assignment of coordinates of an absolute geographic reference system to a

geographic feature”3.

The ScanWorlds coordinate system is based on the scaner's default coordinate system, unless the

scaner was set over known points and these points were imported into ScanControl.

Geo-referencing of scaned data can be defined as a process of transforming the 3D coordinate

vector of the laser sensor frame (S-frame) to the 3D coordinate vector of a mapping frame (m-

frame) in which the results are required. The m-frame can be any earth-fixed coordinate system

such as curvilinear geodetic coordinates (latitude, longitude, and height), UTM, or 3TM

coordinates (Charles et al, 2009).

3 http://www.anzlic.org.au/glossary_terms.html [Retrieved March 23, 2013]

-17-

To geo-reference a given scene, one first needs to establish control points, input the known

geographic coordinates of these control points, e.g. total station measurement, choose the

coordinate system and other projection parameters and then minimize residuals. Residuals are the

differences between the actual coordinates of the control points and the coordinates predicted by

the geographic model created using the control points. They provide a method of determining the

level of accuracy of the geo-referencing process.

3.2 Overview of Total Station

In this thesis Leica 1201 total station (see Fig.3.2) was used. The total station is a surveying

instrument that combines the angle measuring capabilities of theodolite with an electronic

distance measurement (EDM) to determine horizontal angle, vertical angle and slope distance to

the particular point.

Fig.3.2: Leica 1201 Total Station

Coordinates of an unknown point relative to a known coordinate can be determined using the

total station as long as a direct line of sight can be established between the two points. Angles

and distances are measured from the total station to points under survey, and the coordinates (X,

Y, and Z or northing, easting and elevation) of surveyed points relative to the total station

position are calculated using trigonometry and triangulation. To determine an absolute location, a

total station requires line of sight observations and must be set up over a known point or with line

of sight to two or more points with known location4.

Total stations can be manually adjusted or have motors that drive their telescopes very

accurately. The most sophisticated total stations can be operated remotely and continuously at

various levels of automation.

4 http://en.wikipedia.org/wiki/Total_station [Retrieved on March 18, 2013]

-18-

According to Leica geosystem recommendation5, in order to get accurate and precise

measurements in the daily work, it is important:

To check and adjust the instrument from time to time.

To take high precision measurements during the check and adjust procedures.

To measure targets in two faces. Some of the instrument errors are eliminated by

averaging the angles from both faces.

When measurements are being made using the laser EDM, the results may be influenced by

objects passing between the EDM and the target. For example, if the intended target is the

surface of a road, but a vehicle passes between the total station and the target surface, the result is

the distance to the vehicle, not to the road surface.

Instruments equipped with an ATR (Automatic Target Recognition) sensor permit automatic

angle and distance measurements to prisms. The prism is sighted with the optical sight. After

initiating a distance measurement, the instrument sights and centers the prism automatically.

Vertical and horizontal angles and slope distance are measured to the center of the prism and

coordinates of the target calculated automatically.

Using Leica 1200+ instruments, the operator does not have to look through the telescope to align

the prism or a target because of the ATR. This has a number of advantages over a manually

pointed system, since a motorized total station can aim and point quicker, and achieve better

precision (Leica 1200+ TS manual).

3.2.1 Measurement accuracy

Total station measurements are affected by changes in temperature, pressure and relative

humidity, but it can be corrected for atmospheric effects by inputting changes in temperature,

pressure and relative humidity. Shock and stress result in deviations of the correct measurement

as a result decreases the measurement accuracy. Beam interruptions, severe heat shimmer and

moving objects within the beam path can also result in deviations of the specified accuracy by

the manufacture as specified in Table 3.2. It is therefore important to check and adjust the

instrument before measurement.

The accuracy with which the position of a prism can be determined with Automatic Target

Recognition (ATR) depends on several factors such as internal ATR accuracy, instrument angle

accuracy, prism type, selected EDM measuring program and the external measuring conditions.

The ATR has a basic standard deviation level of ± 1 mm but above a certain distance, the

instrument angle accuracy predominates and takes over the standard deviation of the ATR

manual. Leica 1201 total station instruments have standard deviation of 0.3 mgon in both angles

which affect the quality of measurement (Leica 1200+ TPS manual). Typical Leica 1200+

instrument accuracy (horizontal and vertical angles) stated by the manuafacturer are given in the

Table 3.2.

5 http://hds.leicageosystems.com/en/5940.htm [Retrieved on March 20, 2013]

-19-

Table 3.2: Angle measurement accuracy

Type of instrumnt Standared devation (Horizontal and Vertical angles)

[arcsecond] [mgon]

1201+ 1 0.3

1202+ 2 0.6

1203+ 3 1.0

1205+ 5 1.5

Using different prisms other than the intended prism may cause also deviations and therefore it is

important to use a Leica circular prism as the intended target.

3.2.2 Measurement Errors

Some errors, those associated with the instrument, can be eliminated or at least reduced with two

face measurement. Table 3.3 shows instrumental errors which influence both horizontal and

vertical angles, and their adjustment method.

Table 3.3: Angle errors and their adjustment.

Instrument error Affects Hz

angle

Affects V

angle

Eliminated with two

face measurement

Corrected with

instrument calibration

Line of sight error Yes No Yes Yes

Tilting axis error Yes Yes Yes Yes

Compensator errors Yes Yes No Yes

V-index error Yes Yes Yes Yes

Collimation axis error (line of sight error) affects the horizontal angle to be deviated and resulting

in poor accuracy measurement. This axial error is caused when the line of sight (see Fig.3.3) is not

perpendicular to the tilting axis. It affects all horizontal circle readings and increases with steep

sightings, but this effect can be corrected by taking average of two face measurement in two

rounds. For single face measurements, an on-board calibration function is used to determine

collimation errors, the deviation between the actual line of sight and a line perpendicular to the

tilting axis.

Vertical axis error (tilting axis error) errors occur when the titling axis of the total station is not

perpendicular to its vertical axis. This has no effect on sightings taken when the telescope is

horizontal, but introduces errors into horizontal circle readings when the telescope is tilted,

especially for steep sightings. As with horizontal collimation error, this error is eliminated by two

face measurements.

-20-

Fig.3.3: Collimation errors

Compensator index error: errors caused by not leveling a theodolite or total station carefully and

then cannot be eliminated by taking two face measurements. If the total station is fitted with a

compensator it will measure residual tilts of the instrument and will apply corrections to the

horizontal and vertical angles for these.

Vertical Collimation (vertical index) error: a vertical collimation error occurs if the 0o

to 180o line

in the vertical circle does not coincide with the vertical axis. This zero point error is present in all

vertical circle readings and like the horizontal collimation error it is eliminated by taking two

face measurements.

3.2.3 Mode of distance measurement

Measuring with reflector (IR mode)

EDM instruments send a light wave to a reflector and by measuring the phase difference required

in returning the reflected light wave to its source, it computes the distance. Using TS 1201 the

shortest measuring distance is 1.5 m. but, below this limit, there is no possibility to measure. The

specified ranges of different prisms presented in Table 3.4.

Table 3.4: Range limit based on atmospheric condition

Reflector Range A [m] Range B [m] Range C [m]

Standard prism 1800 3000 3500

3600 prism (GRZ4, GRZ 122) 800 1500 2000

3600 Mini prism (GRZ 101) 450 800 1000

Mini prism (GMP101) 800 2600 3300

Three sets of atmospheric conditions:

A: Strong haze, visibility 5 km; or strong sunlight, severe heat shimmer

-21-

B: Light haze, visibility about 20 km; or moderate sunlight, slight heat shimmer

C: Overcast, no haze, visibility about 40 km; no heat shimmer

Accuracy of standard prism distance measurement depends on the type of reflector and the

measuring mode used as indicated in Table 3.5.

Table 3.5: Accuracy of measurements to standard prism

EDM measuring mode Std dev. Standard prism Measremnt time [s]

Standard 1 mm + 1.5 ppm 2.4

Fast 3 mm + 1.5 ppm 0.8

Reflectorless EDM

Distance measurement without reflector (RL mode) is applicable in inaccessible locations such

as building corners, busy highways, top of light pole, etc. Table 3.6 shows distance accuracy in

RL mode. The accuracy depends on the distance between total station and the target to be

measured. The shorter distance the better accuracy.

Table 3.6: Distance accuracy in RL mode

Distance Standard deviation Measuring time, typical [s]

< 500 m 2 mm + 2 ppm 3 – 6

> 500 m 4 mm + 2 ppm 3 - 6

Both reflector (IR) and reflector less (RL) mode measurements have their own advantage and

disadvantages. Their pros and cons are stated in Table 3.7.

Table 3.7: Comparison of IR and RL mode

IR cons IR pros

A person needed for the reflector Can be measured longer distances

Inaccurate for inside corner measurements Faster than reflector less

Measurements are difficult in busy highways,

top of buildings, sites under construction

Better precision than reflector less

RL cons RL pros

Good accuracy only for shorter distances No need person for reflector

Less accurate and slower Can measure inaccessible locations

3.3 Overview of GPS

A GPS receiver (see Fig.3.4) measures the incoming phase of the satellite signals to millimeter

precision. However as the satellite signals propagate from satellites to receivers they pass and are

affected by the atmosphere. The atmosphere that influences the incoming signal consists of the

-22-

ionosphere and troposphere. Disturbance in the atmosphere cause degradation in the accuracy of

the observations (GPS 500 user manual).

GPS surveying is a differential method; a baseline is observed and computed between two

receivers. When the two receivers observe the same set of satellites simultaneously, most of the

atmospheric effects are canceled out. The shorter the baseline is the more these effects will be

reduced, as more likely it is that the atmosphere through which the signal passes to the two

receivers will be identical.

Fig.3.4: GPS receiver

Baseline precision depends on various factors including the number of satellites tracked, satellite

geometry, observation time, ephemeris accuracy, ionospheric disturbance, multi path, resolved

ambiguities, etc.

3.3.1 Real Time Kinematics (RTK)

Real time kinematics data collection uses differential GPS corrections broadcast by a base

receiver to solve for coordinates at a rover receiver in real time. There are several ways to

transmit a correction signal from the base station to mobile station. The most popular way to

achieve real-time transmission is radio communication. The accuracy of the resulting range

measurement depends on the number of satellites in view, resolved ambiguities, satellite

geometry, etc.

RTK mode for geodetic measurements is very fast method for surveying and results are available

immediately, no need for additional data processing afterwards since correction are made from

the base station during the measurement through radio communication (Kostov, 2011).

3.3.2 Comparison of Total Station and GPS

Despite many advantages, surveying using total stations or GPS has disadvantages. Surveying

with a total station, unlike GPS, is not disadvantaged by overhead obstructions but, it is restricted

to measurements between inter-visible points. Often control points are located distant to the

survey area, and traversing with a total station to propagate the control is a time consuming task.

-23-

For this reason, GPS is used to bring control to the survey site through before continuing the

survey with a total station in areas that limit the use of GPS. Table 3.8 shows their advantage and

disadvantages.

GPS can measure points without any line of sight requirement. Since total stations work on the

principle of signal reflection, line of sight must be there between total station and prism reflector.

This makes GPS more effective tool for control point establishment. However, GPS cannot be

used in areas with lot of trees, high rise buildings because of satellite signal interference6.

Table 3.8: Comparison of GPS and total station

Total station GPS

Indirect acquisition of 3D coordinates Direct acquisition of 3D coordinates

Both horizontal and vertical accuracies are

comparable

The horizontal accuracy is better than the vertical

accuracy

The accuracy depends on the distance, angle and

the used prism

The accuracy depends on the satellite availability,

atmospheric effect, satellite geometry, multipath

More precise than GPS Less precise than total station

Satellite independent Satellite dependent

Needed inter-visibility between the instrument and

the prism

Visibility is not needed

Day time data collection Day or night time data collection

3.3.3 Comparison of Total Station and Laser Scaner

A laser scaner is a surveying instrument that determines a three dimensional coordinates of a

given scene in the form of point cloud. Those point clouds represent the position of an object in

3D. Individual points can be compared with points measured by a total station. Their advantages

and disadvantages are presented in Table 3.9.

Table 3.9: Comparison of laser scaner and total station

Laser scanner Total station

Dense information along homogenous surface Single measurement (angles and distance)

of a point

Day or night data collection Day time data collection

Direct acquisition of 3D coordinates Indirect acquisition of 3D coordinates

The vertical accuracy is better than the horizontal

accuracy

Both angles have comparable accuracy

The accuracy depends on the angle and distance

from the facade

The accuracy depends on the distance,

prism used

Heavier to transport Easier to transport

6 http://totalstation.org/gps [Retrived on April 15, 2013]

-24-

Summary of the three methods presented in Table 3.10 for comparison depending on their

advantages and disadvantages

Table 3.10: Conclusion of pros and cons of Total station GPS and Laser scaner

Leica total station 1201 Leica GPS 1201 Leica HDS 2500

Local precision is high (1-2mm

range)

Real time GNSS is (1-2cm)

horizontally and 2-3cm

vertically

Local precision is high, +/-4mm for

range and +/- 60 micro radian for

angle measurements

Uses accurate distance meters

and angle encoders to measure

position to a nearby reflector

GNSS is relaying on satellites

that are approximately 20,000

km away to compute the

rovers’ position.

Has a maximum 40° x 40° field-of-

view. Its 360° x 195° pan & tilt mount

and dual internal rotating mirrors

enable it to be deployed in virtually

any orientation

Provides local coordinates Provides global coordinates Provides local coordinates

Flexibility: used in indoors and

outdoors. Its accuracy is not

degraded by trees blocking or

ionospheric effects.

Used in outdoors GNSS is

not limited to the line of

sight, not weather dependent,

not relay on local land marks

Used in indoors and outdoors. Its

accuracy is not degraded with trees

but it is weather dependent, doesn’t

work below -60c

Weight: lighter than TLS Somewhat heavier than TS Heavier than TS and GPS

Day time data collection Day or night data collection Day or night data collection

Indirect acquisition of

coordinates

Direct acquisition of

coordinates

Direct acquisition of coordinates

3.4 Error analysis

Error is the difference between a measured or calculated and the established value of a quantity.

In the case of this thesis the established value is the values determined through reference network

that controls the detailed survey.

3.4.1 Measurement Errors

There are three types of errors: systematic errors, gross errors and random errors.

Systematic errors are those errors which follow certain physical or mathematical rules. These

kinds of errors are: calibration errors, tension in analogue meters, ambient temperature, etc.

Those errors can be corrected by applying correction factors, calibrating instruments and

selecting suitable instruments.

In most cases gross errors can be caused by human mistakes such as carelessness. The instrument

may be good and may not give any error but still the measurement may go wrong due to the

operator. Those errors do not follow any physical or statistical rules. This can be corrected by

carefulness during the measurement and two face measurements can also detect gross errors.

-25-

Examples of those kinds of errors are: taking wrong readings, wrong recording of instrument or

target height, reading with parallax error, etc.

Random errors are errors in measurement that lead to measured values being inconsistent when

repeated measurements are performed. Those errors are random and affect the measurements in

non-systematic way. Random errors can be caused by instrument errors, human factors, physical

environment, etc. and they can be improved when frequency of measurement is increased, i.e.,

the same parameter is to be measured more often.

Errors in measurements stem from three sources: personal, instrumental, and natural. Personal

errors are caused by the physical limitations of the human senses of sight and touch. An example

of a personal error is an error in the measured value of a horizontal angle, caused by the inability

to hold a range pole perfectly in the direction of the plumb line. Personal errors can be

systematic, random or gross errors. Personal systematic errors are caused by an observer

tendency to react the same way under the same conditions. When there is no such tendency, the

personal errors are considered to be random. When personal mistakes such as; recording 69o

instead of 96o during measurement are gross errors. Instrumental errors are caused by

imperfections in the design, construction, and adjustment of instruments and other equipment.

Instruments can be calibrated to overcome these imperfections. Natural errors result from natural

physical conditions such as atmospheric pressure, temperature, humidity, gravity, wind, and

atmospheric refraction.

3.4.2 Accuracy

Field observations and the resulting measurement are never exact. Any observation can contain

various types of errors. Often some of these errors are known and can be eliminated or at least

reduced by applying appropriate corrections. However, even after all known errors are

eliminated, a measurement will still be in error by some unknown value. To minimize the effect

of errors and maximize the accuracy of the final result, the surveyor has to use utmost care in

making the observations. However, a measurement is never exact, regardless of the precision of

the observations.

Accuracy is the degree of conformity with a standard or accepted value. Accuracy relates to the

quality of the result. The standards used to determine accuracy can be:

– An exact known value, such as the sum of the three interior angles of a plane triangle is

180°.

– A value of a conventional unit as defined by a physical representation thereof, such as the

international meter.

– A survey determined or established by superior methods and deemed sufficiently near the

ideal or true value to be held constant for the control of detail survey.

-26-

The accuracy of a field survey depends directly upon the precision of the survey. Therefore, all

measurements and results should be quoted in terms that are commensurate with the precision

used to attain them. Similarly, all surveys must be performed with a precision that ensures that

the desired accuracy is attained. Although they are known to be not exact, established control

points are deemed of sufficient accuracy to be the control for all other detail surveys.

3.4.3 Precision

Precision is the ability to repeat the same measurement. It is a measure of the uniformity or

reproducibility of the result. Precision is different from accuracy in that it relates repeatability of

the measurements made. In short a measurement is precise if it obtains similar results with

repeated measurements, while accuracy is the closeness to the established value.

3.4.4 Checking accuracy

It is true that any measurement would not be free from errors. In most cases gross errors may

happen in a measurement and therefore the accuracy of the measurement needs to be checked in

order to avoid the gross errors. There are a lot of accuracy checking mechanisms, for instance,

through two face measurement, adjustment, etc. Using these mechanisms, gross errors can be

detected. As Csanyi et al, (2007) stated out, small magnitude errors of each individual

measurement may affect the quality of the final result by considerable large amount. Therefore,

the final result may depend on the quality achieved from each individual measurement.

3.4.5 Quality Control

The term quality control (QC) refers to the efforts and procedures that researchers put in place to

ensure the quality and accuracy of data being collected using the methodologies chosen for a

particular study (Roe, D., 2008).

Quality control measure verifies the accuracy of the surveyed data by checking its compatibility

with an independently surveyed data. For instance: in the comparison of TS and TPS, laser

scaner targets were extracted from the range of scaning. The coordinates of the extracted targets

are then compared with the independently TS surveyed coordinates using RMS analysis. Thus,

the total station measurement controls quality of the TLS extracted points. As per Habib et al,

(1999), the resulting RMS value is a measure of the external and absolute quality of the scaned

derived surface.

-27-

4 METHODOLOGIES

4.1 Establishing Reference Network

In order to evaluate the accuracy and precision of the surveyed data, primary it has been

established a network of control points which can serve as a reference for comparison with RTK

and TLS measurement. The reference network was established fourteen control points using a

Leica 1201 total station. To determine the network with high precision, measurements have been

taken in two faces with two rounds. Four points of the reference network were also measured

with static GPS in order to transform the datum from the local coordinate system to the required

coordinate system, SWEREF 99. Thus, this network served as a reference value. The precision of

the remaining RTK and TLS measurements were evaluated depending on this reference value.

Therefore, to accomplish the objectives of this project, data were collected from field

measurement. The field measurements were taken using three different surveying instruments: -

Global Positioning System (GPS), laser scaner (LS) and total station (TS). To eliminate

instrumental errors such as line of sight errors, tilting axis errors and vertical index errors (see

Table 3.3), two face measurements were taken. Since the coordinates determined with total

station are provided in local coordinate system, static GPS measurement was needed to transform

the datum to SWEREF 99. Then, precision of the network has been obtained from network

adjustment and verified for if there have been gross errors were occurred. Detail measurements

(RTK on the network and, TLS and TS on the façade) were taken five times to evaluate the

precision of the measurement. Finally, accuracy and precision of the detail measurements were

tested by RMS and standard deviation analysis respectively.

4.2 Evaluation of Accuracy and Precision

To evaluate the accuracy and precision of the measurement, RMS and standard deviation of the

individual measurements were computed. RMS (root mean square error) is a measure of

accuracy of the individual measurement. It can be computed from the deviations between true

and measured values. True value of the measured quantity is the value which was determined

with significantly higher precision. In this project the coordinates of the reference network were

considered as ‘true’ which is determined in 1mm level. RMS was computed using the following

formula:

(4.1)

Where: is the established value, is individual measurement and is the number of

measurements.

-28-

Standard deviation is a measure of variations of the repeated measurement, i.e. of the precision of

each individual observation. It can be computed from the mean values of the individual

measurement and the individual measurement. Standard deviation is computed using the

following formula.

(4.2)

Where: is true or established value, is individual measurement, is mean value of the

measurements and is number of measurements.

4.2.1 Choosing suitable control points for the network and detail survey

Reconnaissance of the project area was the first step in the establishment of control network and

followed by marking fourteen control points which are visible each other. Those control points

were also suitable for satellite visibility, because RTK method was needed to compare with the

TS control points. The points are marked with nails for sustainability reason. The project area

was close to L building in the campus of KTH, Stockholm, Sweden (see Fig 4.1).

4.2.2 Setting up targets for laser scaning

In order to compare the results from total station and laser scaner, 21 target points were chosen at

the North Eastern façade of the L building. Six black and white target papers were marked as

control points for the registration of ScanWorlds. Those target points were also measured with

total station. There are requirements to be fulfilled when choosing black and white targets. As

Quintero et al, (2008) stated out, not only is the station position important, the positioning of the

targets carries equal importance. And so, it is important to note that:

targets be widely separated;

targets have different heights;

as few targets as possible on one single line;

4.2.3 Detail survey

Once the reference network and the targets for detail measurement were established, the next step

was taking the detail survey. RTK measurement was taken on the reference network to compare

the result with total station measurement, and measurements from laser scaning and total station

on the façade of L building were taken and the results were compared. In order to evaluate the

precision of the measurements, all control points and targets points were measured five times.

During all measurements the time required was recorded for comparison.

Total station

In order to determine and compare accuracy, precision and time expenditure of the this method,

the façade of L building with black and white paper targets and corners of windows were

-29-

surveyed five times with the total station. The data was processed in Geo and then, the obtained

coordinates of the facade targets were used as constraint during the registration and geo-

referencing processes. And time expenditure was also recorded both for field measurement and

processing.

Laser Scaning

The façade of the north eastern of the L building was scaned five times with laser scaner from

five different views. The scaned ranges were between 6 m to 9 m. Captured point clouds were

registered and geo-referenced with precisely determined total station data. Time expenditure for

scaning and processing were recorded, analyzed and presented in Table 5.13.

GPS RTK (Real Time Kinematics)

The RTK method was performed to compare accuracy of the network with total station

measurements. Using one known coordinate point (DUB) from the adjusted reference network,

RTK was used to measure the remaining 13 points five times with 3D quality reported by the

receiver of less than 9mm. This 3D quality describes the accuracy of the GPS measurement.

Depending on the satellite availability and other sources of errors that affect the GPS

measurement, the magnitude of the 3D quality might be small or large. If there is good satellite

geometry (i.e. satellites scattered around the four quadrants), good satellite visibility and other

GPS errors are small, the 3D quality will be small otherwise it will be large. The 3D quality

( ) can be computed using the formula below (Eq. 4.3):

4.3

Where: is standard deviation of X, Y and Z coordinates

Results of each method were analyzed and compared in order to evaluate the accuracy, precision

and time expenditure.

Project area of the study

The project area is the parking lot close to L building, KTH campus, Stockholm, Sweden

(Fig.4.1). First, reconnaissance of the project area has been performed, and followed by

establishing a network of 14 control points, which have been used as a reference value for the

detail survey. The network has been established using Leica 1201 version total station. Figure 4.1

shows the project area and the reference control points. In the Fig.4.1, points dub1, N1, C1 and H1

were measured also by static GPS.

-30-

Fig.4.1: Reference Control Points

Data processing

Data were processed in the respective software of the instruments. Data from laser scanning was

processed in Cyclone; data from the total station processed in Geo and data from GPS processed

in Leica geo office. Registration and geo-referencing of the point cloud was performed to

combine ScanWorlds together in one coordinate system of the scanner and then transformed to

SWEREF 99.

As a matter of human limitations, imperfect instruments, unfavorable physical conditions and

improper measurement routines, which together define the measurement condition, all

measurement results most likely contain errors. To reduce the measurement errors on the final

results one need to improve the overall condition of the measurement using least square

adjustment (Fan, 1997).

Adjustment of the network was performed in Geo software which uses method of least square

adjustment. Least square adjustment is a method of estimating values from a set of observations

by minimizing the sum of the squares of the differences between the observations and the values

to be found.

Least squares method is a classical method which defines the optimal estimate of X (unknown)

by minimizing the sum of the weighted observation residuals squared (Fan, 1997).

(4.4)

-31-

P =

(P1,P2,…,Pn)

Where ith measurement

: residual vector and

: number of observations

In Eq. 4.4 the weight matrix (P) is introduced because the network adjustment was a result of

distances and angular measurements. In the adjustment process both distances and angles have

different weights of a priori standard deviations. Thus, P matrix has been introduced.

Let represent adjusted value and its residual of observation such that

.

Here, is a non-linear function of unknown parameters :

(4.5)

Let denote an approximate value of and its corresponding correction, such

that :

(4.6)

The non-linear equation (4.5) can be expanded by Taylor series and the linearized equation

found:

(4.7)

Where:

(4.8)

According to Fan, (1997), the linear system is:

(4.9)

Where:

-32-

(4.10)

The least square solution of can be written as:

(4.11)

The least squares solutions of the unknown parameter can be computed as:

(4.12)

The least square estimates of the residual, is calculated as:

(4.13)

Then, the posteriori standard deviation of unit weight is computed as:

(4.14)

And the cofactor matrix is computed as:

(4.15)

Hence the variance-covariance matrix becomes:

(4.16)

And the standard deviation of the unknown parameters can be computed from the diagonal

cofactor matrix as:

(4.17)

Here: is the diagonal element of cofactor matrix in (Eq. 4.15)

Gross errors can be detected with standardized residual using the following formula:

(4.18)

Here: is the weight of ith

measurement and is the diagonal matrix of :

(4.19)

Here: is the redundancy and I is the identity matrix

-33-

5 RESULTS AND DISCUSSION

5.1 GPS Baseline Processing

Four control points of the reference network were observed with static measurement for three

hours. In order to transform the observed points from WGS 84 to SWEREF 99, a baseline was

processed from these four control points to SWEPOS station which is Continuously Operating

Reference Station (CORS) that provides Global Navigation Satellite System (GNSS). These data

are consisting of carrier phase and code range measurements in support of three dimensional

positioning. After processing the baselines, the coordinate system was changed in to SWEREF

99. Those coordinates were used as known in the adjustment of the reference network.

The computed coordinates and their standard deviations of the reference network are presented in

the Table 5.1.

Table 5.1: Computed coordinates (m)

Point N E

H

C 6581703.78 153915.061 58.067 0.001 0.000 0.001

H 6581705.635 153894.633 57.700 0.001 0.000 0.001

N 6581725.636 153874.812 57.523 0.001 0.001 0.000

DUB 6581730.261 153920.049 57.029 0.001 0.000 0.001

5.2 Adjustment

Adjustment is an improvement of the measurement, since measurements are not free from errors.

Improvements to observations and coordinates for new points are calculated with various quality

measures such as standardized residuals, standard deviation, redundancy numbers, error ellipses

etc. The reference network was adjusted first with free adjustment in order to eliminate any

contradictions in the fixed points.

5.2.1 Rounds of Measurement

For the sake of eliminating or at least reducing errors emanating from collimation axis error,

vertical axis error, compensator errors (longitudinal and transverse), vertical index errors, two

face measurements with two rounds have been taken. Mean values of the two face measurements

were checked if their differences were below 2 mm for distance and 6 mgon for angles.

Atmospheric corrections were also applied before adjustment.

5.2.2 A priori standard deviation

A priori standard deviations have to be considered in the input observation data. Since the

measurements consisted of distances and angles, they have different weights to be applied in the

adjustment using Eq. (4.4).

-34-

These a priori standard deviations are provided by the manufacturers. For the adjustment of the

reference network, a priori standard deviation (see Fig.5.1) for the distance was 2 mm + 2 ppm,

for the direction 0.6 mgon and standard deviation for the centering error was 1 mm. A priori

standard deviation of the height of instrument was 3 mm.

Fig.5.1: A priori standard deviation

5.2.3 Horizontal Network Adjustment

Planimetric coordinates (N and E) were adjusted with free adjustment with translation and

rotation. This is a type of adjustment when the reference network is adjusted initially as fully free

and is then connected with a transformation. The net fits the known points through the translation

in the N and E axes and a rotation. Table 5.2 shows values before adjustment and after

adjustment. The residual values computed using Eq. (4.13) show the difference between the

adjusted values minus the measured values. The color is controlled by the residual size: Green, if

the residual is less than 1 a priori standard deviation. Black, if the residual is less than 2 a priori

standard deviations and Red if the residual is greater than 2 a priori standard deviations, in this

case the measurement is likely to contain errors. Residuals greater than 2 a priori standard

deviations were checked if serious errors had occurred.

-35-

Table 5.2: Some of adjusted values and their standard deviations

Obs Type Station Object Value A pr.

StdDev

A pr.

SD+C

Residual Adj.

value

StdDev Std

Residual

Direc_n H1 E1 210.57 0 0.004 -0.006 210.566 0.004 -1.222

Length H1 E1 16.579 0.005 0.005 -0.005 16.574 0.001 -0.614

Direc_n H1 G1 223.51 0 0.008 -0.016 223.497 0.009 -1.461

Length H1 G1 8.08 0.005 0.005 -0.002 8.078 0.001 -0.195

Direc_n H1 F1 240.23 0 0.004 -0.005 240.227 0.005 -1.192

Length H1 F1 16.672 0.005 0.005 -0.004 16.669 0.002 -0.408

Direc_n M1 C1 83.857 0 0.001 -0 83.857 0.001 -0.064

Length M1 C1 58.17 0.005 0.005 0.007 58.177 0.002 0.831

Direc_n M1 dub1 53.223 0 0.001 -0.001 53.222 0.001 -0.803

Length M1 dub1 52.586 0.005 0.005 0.002 52.589 0.002 0.239

Direc_n M1 H1 100.01 0 0.002 0.001 100.015 0.001 0.188

Length M1 H1 41.903 0.005 0.005 0.002 41.906 0.002 0.267

Direc_n M1 N1 111.94 0 0.004 0.008 111.949 0.005 1.359

Length M1 N1 14.123 0.005 0.005 -0.004 14.119 0.002 -0.449

Direc_n M1 L1 0 0 0.005 -0.007 399.993 0.007 -1.488

Length M1 L1 12.486 0.005 0.005 -0.002 12.484 0.001 -0.183

Direc_n M1 K1 60.647 0 0.004 -0.008 60.639 0.003 -1.278

Length M1 K1 15.649 0.005 0.005 0.002 15.651 0.001 0.181

Direc_n M1 A1 61.966 0 0.001 0.003 61.97 0.001 2.241

Length M1 A1 62.24 0.005 0.005 0.008 62.248 0.002 0.933

As Table 5.2 shows, the maximum and minimum standard deviation of the individual observation

of the network was 9 gon in direction and 1 mm in distance respectively. This indicates the

measurement was done accurately without gross errors.

Standardized residual values which are computed using Eq. (4.18) are measures of gross errors

detected. The colors are also controlled by the standardized residual size. Green: The standard

residual is less than 1. Black: The standard residual is less than 2. Red: The standard residual is

greater than 2, in which the measurement may be inaccurate and should be checked.

5.2.4 Adjusted coordinates and their standard deviations

Coordinates of all points (Table 5.3) in the reference network were calculated with free

adjustment with translation and rotation. First adjusted as fully free and then connected with the

X and Y coordinate axes transformation. The net fits the known points through the translation in

the X and Y axes and a rotation. Obtained coordinate errors of the network were below 1mm,

maximum standard deviations were 0.9 mm in horizontal and 0.7 mm in height. This indicates

the network was established with high precision and therefore served as a reference value for the

remaining detail measurements. GPS-RTK, laser scaning and total station measurements were

evaluated with reference to the established value.

-36-

Table 5.3: Adjusted coordinates of the reference network

Point N E H

A 6581720.4857 153927.7596 57.6203 0.0008 0.0006 0.0004

C 6581703.7786 153915.0632 58.2209 0.0007 0.0007 0.0004

D 6581697.5697 153911.0546 58.4491 0.0006 0.0007 0.0004

E 6581692.1084 153904.1971 58.5289 0.0010 0.0008 0.0007

F 6581689.1587 153897.1126 58.6269 0.0012 0.0009 0.0007

G 6581698.2412 153897.8686 57.9827 0.0009 0.0009 0.0007

H 6581705.6416 153894.6336 57.7450 0.0006 0.0006 0.0005

I 6581713.1413 153893.4920 57.5835 0.0009 0.0009 0.0007

J 6581723.8830 153889.5021 57.2779 0.0011 0.0010 0.0007

K 6581733.9285 153883.1438 56.9147 0.0006 0.0007 0.0004

L 6581745.9489 153877.6836 56.5120 0.0010 0.0009 0.0007

M 6581738.0219 153868.0418 56.7599 0.0007 0.0007 0.0004

N 6581725.6336 153874.8107 57.2913 0.0009 0.0010 0.0007

DUB 6581730.2577 153920.0470 57.0608 0.0007 0.0006 0.0004

As it is mentioned in the introduction part, the network consisted of fourteen control points.

Fig.5.2 shows the distribution of standard deviation ( ) levels, the colors are controlled by the size

of sigma. The measurement's sigma level corresponds directly to the absolute value of the

standardized residual. The standardized residual 0-1 gives sigma level 1, 1-2 gives sigma level 2,

2-3 gives sigma level 3 and greater than 3 gives sigma level 3+. Therefore, this sigma level

shows quality of the measurement. The least sigma level means the more precise measurement

was done and the large sigma level shows the measurement likely to contain gross errors which

should be verified and rejected.

-37-

Fig.5.2: Graphical view of the reference network

5.2.5 Vertical Network adjustment

For the vertical network (height) adjustment, data from vertical angles and slope distances has

been used. First it was adjusted as fully free and then connected with the transformation. The

network fitted the known points through the translation of Z axis. The vertical accuracy was also

determined below 1 mm level, the maximum and minimum standard deviations were 0.7 mm and

0.4 mm respectively.

.

Fig.5.3: Graphical view of height adjustment

-38-

5.3 Determination of precision and accuracy of RTK

On the reference network, RTK measurements were taken in order to compare with the total

station measurements. Using RTK method, all control points were surveyed five times so as to

evaluate the precision of the measurements. To compute the precision of the repeated

measurement of the reference network, standard deviation formula Eq. (4.2) has been used. Then,

RMS of the RTK measurements were also computed using Eq. (4.1) in order to evaluate how

much the measurements were close to the established value.

As the result shows in Table 5.4, the standard deviations are less than 8 mm in horizontal and they

reach 1.5 cm in vertical coordinate, which indicates that the repeated measurements were quite

close to each other. According to the results obtained by Jonsson et al (2003), the standard

deviations for the horizontal and vertical coordinate are 9 mm and 2 cm respectively. So, by

comparing the author’s result with this thesis result, the precisions of the horizontal and vertical

coordinate are in mm and cm level respectively.

To evaluate how much RTK measurements were close to the established value, RMS of the RTK

measurements were computed (see Table 5.4). This RMS indicates the accuracy of the RTK

measurements of the reference network. Accuracy of the horizontal coordinates ranges between

maximum 9 mm (points A and D) and minimum 2 mm (point E) and accuracy of the vertical

coordinates ranges between maximum 2.2 cm (point M) and minimum 1.1 cm (point I). This

result can be compared with the work of Ehsani et al, (2004), in which, a horizontal accuracy of

1 cm achieved by compensating for atmospheric delay, orbital errors and other variables in GPS

geometry. By comparing the accuracy of horizontal coordinates, they are close to each other. The

thesis results are quite reasonable considering the errors attributed from satellite blocking,

centering error and so on.

Table 5.4: RTK measurement, its RMS and standard deviations

X Y Z X Y Z

Point RTK Mean RMS St.D

A 6581720.487 153927.762 57.616 0.003 0.009 0.012 0.003 0.006 0.015

C 6581703.78 153915.066 58.228 0.007 0.008 0.017 0.007 0.006 0.008

D 6581697.574 153911.055 58.441 0.009 0.005 0.017 0.003 0.004 0.006

E 6581692.109 153904.196 58.521 0.005 0.002 0.015 0.004 0.001 0.005

F 6581689.157 153897.11 58.621 0.004 0.006 0.015 0.003 0.001 0.004

G 6581698.243 153897.867 57.977 0.003 0.004 0.012 0.001 0.003 0.006

H 6581705.645 153894.637 57.738 0.007 0.008 0.015 0.004 0.004 0.007

I 6581713.144 153893.493 57.579 0.007 0.008 0.011 0.005 0.008 0.005

J 6581723.884 153889.504 57.271 0.005 0.007 0.013 0.005 0.005 0.003

K 6581733.927 153883.147 56.907 0.003 0.006 0.015 0.001 0.001 0.002

L 6581745.959 153877.687 56.506 0.008 0.007 0.013 0.007 0.002 0.004

M 6581738.026 153868.045 56.749 0.008 0.007 0.022 0.002 0.003 0.001

N 6581725.636 153874.814 57.285 0.005 0.006 0.012 0.003 0.003 0.006

-39-

In order to check if there were significant differences between total station and RTK results, the

difference between the total station and the RTK measurements were computed. The difference

was computed using the mean values of the measurements. According to the obtained result, it

has been proved that there were no significant differences between the two measurements. As

indicated in Table 5.5, the maximum difference between TS and RTK is 1 cm in horizontal and

1.1 cm in vertical coordinates. The remaining coordinates are below this level Thus, by

comparing the mean coordinates of the RTK measurements with that of total station, their

measurement quality has been verified. Therefore, it can be concluded that the measurements

were precise and accurate. This can be compared with the standard deviations of the difference

between the total station and the RTK measurements.

Table 5.5 The difference between TS and RTK mean

TS Mean RTK Mean TS-

RTK

Point N E H N E H

A 6581720.486 153927.76 57.6203 6581720.487 153927.762 57.616 -0.001 -0.002 0.004

C 6581703.779 153915.063 58.2209 6581703.78 153915.066 58.228 -0.001 -0.003 -0.007

D 6581697.57 153911.055 58.4491 6581697.574 153911.055 58.441 -0.004 0.000 0.008

E 6581692.108 153904.197 58.5289 6581692.109 153904.196 58.521 -0.001 0.001 0.008

F 6581689.159 153897.113 58.6269 6581689.157 153897.11 58.621 0.002 0.003 0.006

G 6581698.241 153897.869 57.9827 6581698.243 153897.867 57.977 -0.002 0.002 0.006

H 6581705.642 153894.634 57.745 6581705.645 153894.637 57.738 -0.003 -0.003 0.007

I 6581713.141 153893.492 57.5835 6581713.144 153893.493 57.579 -0.003 -0.001 0.005

J 6581723.883 153889.502 57.2779 6581723.884 153889.504 57.271 -0.001 -0.002 0.007

K 6581733.929 153883.144 56.9147 6581733.927 153883.147 56.907 0.002 -0.003 0.008

L 6581745.949 153877.684 56.512 6581745.959 153877.687 56.506 -0.010 -0.003 0.006

M 6581738.022 153868.042 56.7599 6581738.026 153868.045 56.749 -0.004 -0.003 0.011

N 6581725.634 153874.811 57.2913 6581725.636 153874.814 57.285 -0.002 -0.003 0.006

DUB 6581730.258 153920.047 57.0608 6581730.258 153920.047 57.061 0.000 0.000 0.000

In addition, the standard deviation of the difference between total station and RTK measurements

were calculated using (Eq. 5.2) and compared the result with their coordinate differences. Table

5.6 presents the standard deviation of the difference between total station and RTK. The result

shows maximum standard deviation difference of 9 mm horizontally and 1.7 cm vertically. Here

the RMS of the total station was not computed because there was no reference value for it.

-40-

Table 5.6: Comparison of standard deviations between TS and RTK

TS

RTK

Diff

Point

A1 0.001 0.001 0.000 0.007 0.006 0.012 0.007 0.006 0.012

C1 0.001 0.001 0.000 0.009 0.005 0.014 0.009 0.005 0.014

D1 0.001 0.001 0.000 0.007 0.008 0.006 0.007 0.008 0.006

E1 0.001 0.001 0.001 0.008 0.006 0.014 0.008 0.006 0.014

F1 0.001 0.001 0.001 0.007 0.008 0.012 0.007 0.008 0.012

G1 0.001 0.001 0.001 0.005 0.008 0.011 0.005 0.008 0.011

H1 0.001 0.001 0.001 0.005 0.009 0.017 0.005 0.009 0.017

I1 0.001 0.001 0.001 0.009 0.001 0.013 0.009 0.001 0.013

J1 0.001 0.001 0.001 0.009 0.007 0.014 0.009 0.007 0.014

K1 0.001 0.001 0.000 0.005 0.008 0.013 0.005 0.008 0.013

L1 0.001 0.001 0.001 0.005 0.006 0.014 0.005 0.006 0.014

M1 0.001 0.001 0.000 0.007 0.007 0.012 0.007 0.007 0.012

N1 0.001 0.001 0.001 0.004 0.006 0.013 0.004 0.006 0.013

Reliability of the measurements was tested through confidence level. Since the number of

measurement is small (n=5), t-distribution was used to compute the confidence interval. t-

distribution is a type of probability distribution that resembles normal distribution but for smaller

sample size. t-distribution is a bell shaped probability distribution with heavier tail producing

values that fall far from the mean. Using 95% confidence level, reliability of the measurements

have been verified.

The range of values within which the true value should lie for a given probability is required.

This range is called the confidence interval, its bounds called the confidence limits. Confidence

limits can be established for that stated probability from the standard deviation for a set of

observations. Statistical tables are available for this purpose. A figure of 95% frequently chosen

implies that nineteen times out of twenty the true value will lie within the computed limits. The

presence of a very large error in a set of normally distributed errors, suggests an occurrence to

the contrary and such an observation can be rejected if the residual error is larger than three times

the standard deviation.

The confidence interval for the difference between total station (TP) and real time kinematics

(RTK) has been calculated as:

(5.1)

(5.2)

-41-

(5.3)

So, the confidence interval is:

(5.4)

Where; is the difference between TPS and RTK coordinates, i = (N, E, H)

coordinates

is the standard deviation of

, P = coordinate point and k is t-

score value calculated from confidence level and degree of freedom.

Table 5.7 presents the confidence interval for the difference between coordinates of RTK and

TPS. Since the theoretical difference between their coordinates is zero, the confidence interval is

[ , where k is t-score value calculated from confidence level, 95% and degree of

freedom (k=2.776). Depending on this requirement, 95% of the points should be inside this

confidence interval limit.

Table 5.7: Confidence interval limits and coordinates difference between TPS and RTK

Confidence interval limit and errors difference between TPS and RTK

Abs[

]

Point N E H N E H

A1 0.0196 0.0178 0.0339 0.001 0.006 0.012

C1 0.0251 0.0146 0.0394 0.006 0.010 0.016

D1 0.0195 0.0212 0.0175 0.012 0.000 0.015

E1 0.0218 0.0163 0.0378 0.008 0.001 0.012

F1 0.0197 0.0229 0.0339 0.015 0.005 0.011

G1 0.0147 0.0212 0.0312 0.005 0.001 0.005

H1 0.0129 0.0245 0.0467 0.013 0.003 0.004

I1 0.0240 0.0039 0.0367 0.018 0.007 0.005

J1 0.0241 0.0191 0.0384 0.011 0.002 0.011

K1 0.0140 0.0217 0.0356 0.026 0.023 0.018

L1 0.0142 0.0168 0.0400 0.010 0.003 0.016

M1 0.0195 0.0201 0.0339 0.014 0.006 0.003

N1 0.0119 0.0180 0.0361 0.002 0.003 0.005

From the results calculated, only one point (K) exceeds the confidence interval limit. Therefore,

point K should be rejected. The rest points have laid within the confidence interval limit, which

accounts 92%.

-42-

5.4 Registration and Geo referencing

In order to register the ScanWorlds, five targets (black and white papers) have been used on the

façade of L-building. Then using the HDS2500 scaner five ScanWorlds were scaned with more

than 80% overlap to each other. As it is described in sub-section 2.1.1, the registration refers

merging multiple scans each other in a correct, relative 3D geometry within a single coordinate

system. Since those common target points were also measured with the total station five times in

two faces, the coordinates have been transformed into SWEREF 99 through geo-referencing. The

result shows coordinates of all points were calculated with mean absolute error of constraints less

than 9 mm. Finally, the mean values of the coordinates of five ScanWorlds were computed in

order to evaluate the precision of the laser scaning measurement. The difference between laser

scaner measurement and total station measurement were determined and compared in the same

procedure as RTK and TS. After ScanWorlds have been registered, TS measured points were

imported so as to transform the scanner’s coordinate system in to the required coordinate system,

SWEREF 99.

Fig.5.4 shows all geo-referenced points on the façade of L building. For the sake of simplicity

understandability, the author used black and white (BW) targets, corners of the windows mirror

(WM) and edges of the black and white target papers (P). These coordinates were determined

with mean absolute error of less than 9 mm.

Fig.5.4: Geo-referenced points

In order to eliminate gross errors from the measurement, two big errors (greater than 1 cm level)

in the measurement were identified and rejected from the registration Fig.5.5 shows accuracy of

registered ScanWorlds. In addition, each error vector of the targets was also checked individually

if gross errors had been occurred.

-43-

Fig.5.5: Registered and geo-referenced ScanWorlds

5.5 Comparison of Laser Scaner and Total Station result

Results from laser scaner and total station measurements are quite similar in both horizontal and

vertical coordinates. As it is described in section 4.5, ScanWorlds were scaned five times to

compute the precision of the measurements. Registration needs at least two ScanWorlds at a time

so as to provide the ScanWorlds in the same coordinate system. So, by forming five pairs of

ScanWorlds from the five measurements, five registrations have been performed using the five

black-white targets as constraint. After performing five registrations, standard deviations of the

measurement were calculated as it is shown in Table 5. 8. It shows that mm level precision has

been achieved for all coordinates.

In the same way, mm level precision has been achieved with the total station also (see Table 5.9).

Even if their magnitude is within mm level, still it could possible to achieve better precision with

the total station.

-44-

Table 5. 8: Standard deviations of the registered TLS

Extracted points from Terrestrial Laser Scaning

Point N E H

BW10 6581693.863 153909.234 61.651 0.002 0.005 0.001

BW11 6581693.247 153908.272 60.000 0.005 0.004 0.012

BW12 6581692.756 153907.503 61.529 0.005 0.005 0.002

BW13 6581692.385 153906.925 60.245 0.003 0.006 0.006

BW15 6581691.992 153906.315 59.442 0.002 0.003 0.003

P101 6581693.915 153909.322 61.803 0.004 0.007 0.004

P102 6581693.938 153909.322 61.505 0.009 0.007 0.006

P103 6581693.808 153909.146 61.508 0.006 0.003 0.009

P104 6581693.806 153909.147 61.803 0.006 0.013 0.006

P111 6581693.296 153908.357 60.143 0.006 0.008 0.004

P112 6581693.296 153908.363 59.852 0.006 0.010 0.003

P113 6581693.186 153908.188 59.845 0.004 0.003 0.008

P114 6581693.186 153908.187 60.143 0.006 0.004 0.003

RT1 6581691.413 153905.406 60.974 0.003 0.006 0.005

RT2 6581691.386 153905.364 61.049 0.006 0.007 0.003

RT3 6581691.435 153905.447 61.048 0.006 0.003 0.003

RT4 6581691.415 153905.406 61.005 0.006 0.006 0.011

WM1 6581693.392 153908.594 60.445 0.007 0.008 0.006

WM2 6581693.386 153908.608 61.236 0.004 0.006 0.004

WM3 6581693.794 153909.284 61.225 0.008 0.016 0.006

WM4 6581693.804 153909.266 60.454 0.005 0.004 0.005

In order to compare the results of the laser scaning measurement with that of total station, five

measurements were taken with total station on the façade of L-building. The black and white

(BW) targets, the edges of these BW targets and the corners of windows mirrors (WM) were

measured on the façade. Then the mean values of the measurements have been determined to

compute its precision. Table 5.9 presents the mean values of five measurements and their standard

deviations. The standard deviations are quite small relative to the laser scaninig measurements.

-45-

Table 5.9: Standard deviations of TS measurements

Total Station points

N E H

BW10 6581693.861 153909.232 61.652 0.002 0.008 0.001

BW11 6581693.246 153908.270 59.993 0.001 0.001 0.001

BW12 6581692.755 153907.501 61.528 0.004 0.002 0.001

BW13 6581692.387 153906.927 60.246 0.001 0.002 0.001

BW15 6581691.991 153906.316 59.443 0.001 0.003 0.001

P101 6581693.916 153909.320 61.802 0.002 0.001 0.002

P102 6581693.926 153909.320 61.507 0.006 0.001 0.002

P103 6581693.806 153909.146 61.505 0.002 0.003 0.001

P104 6581693.805 153909.144 61.801 0.002 0.002 0.001

P111 6581693.299 153908.359 60.144 0.001 0.001 0.001

P112 6581693.298 153908.365 59.851 0.001 0.002 0.002

P113 6581693.187 153908.188 59.844 0.001 0.003 0.001

P114 6581693.187 153908.186 60.142 0.001 0.003 0.001

RT1 6581691.413 153905.408 60.972 0.003 0.002 0.001

RT2 6581691.387 153905.367 61.049 0.002 0.001 0.001

RT3 6581691.439 153905.447 61.049 0.001 0.003 0.001

RT4 6581691.414 153905.407 61.003 0.006 0.001 0.001

WM1 6581693.394 153908.596 60.448 0.003 0.003 0.003

WM2 6581693.385 153908.606 61.234 0.003 0.003 0.003

WM3 6581693.796 153909.288 61.228 0.003 0.002 0.004

WM4 6581693.805 153909.265 60.453 0.004 0.003 0.003

-46-

Table 5.10: Standard deviation and RMS of TLS measurement

TLS Mean values RMS Std D

N E H N E H

BW10 6581693.863 153909.234 61.651 0.002 0.002 0.001 0.002 0.005 0.001

BW11 6581693.247 153908.272 60 0.001 0.002 0.007 0.005 0.004 0.012

BW12 6581692.756 153907.503 61.529 0.001 0.002 0.001 0.005 0.005 0.002

BW13 6581692.385 153906.925 60.245 0.002 0.002 0.001 0.003 0.006 0.006

BW15 6581691.992 153906.315 59.442 0.001 0.001 0.001 0.002 0.003 0.003

P101 6581693.915 153909.322 61.803 0.001 0.002 0.001 0.004 0.007 0.004

P102 6581693.938 153909.322 61.505 0.012 0.002 0.002 0.009 0.007 0.006

P103 6581693.808 153909.146 61.508 0.002 0 0.003 0.006 0.003 0.009

P104 6581693.806 153909.147 61.803 0.001 0.003 0.002 0.006 0.013 0.006

P111 6581693.296 153908.357 60.143 0.003 0.002 0.001 0.006 0.008 0.004

P112 6581693.296 153908.363 59.852 0.002 0.002 0.001 0.006 0.01 0.003

P113 6581693.186 153908.188 59.845 0.001 0 0.001 0.004 0.003 0.008

P114 6581693.186 153908.187 60.143 0.001 0.001 0.001 0.006 0.004 0.003

RT1 6581691.413 153905.406 60.974 0.000 0.002 0.002 0.003 0.006 0.005

RT2 6581691.386 153905.364 61.049 0.001 0.003 0 0.006 0.007 0.003

RT3 6581691.435 153905.447 61.048 0.004 0 0.001 0.006 0.003 0.003

RT4 6581691.415 153905.406 61.005 0.001 0.001 0.002 0.006 0.006 0.011

WM1 6581693.392 153908.594 60.445 0.002 0.002 0.003 0.007 0.008 0.006

WM2 6581693.386 153908.608 61.236 0.001 0.002 0.002 0.004 0.006 0.004

WM3 6581693.794 153909.284 61.225 0.002 0.004 0.003 0.008 0.016 0.006

WM4 6581693.804 153909.266 60.454 0.001 0.001 0.001 0.005 0.004 0.005

Table 5.10 shows the quality of terrestrial laser scaning measurements in the form of precision and

accuracy. The accuracy is determined with 2 mm for both horizontal and vertical coordinates.

The difference between the established and measured values theoretically should be zero, which

is the center of the t-distribution graph. To calculate the upper and lower interval limit, constant k

could be computed from confidence level and degree of freedom. Since the number of

measurements is n=5, degree of freedom will be n-1, which is 4. And k was computed in excel

using the function [tinv(0.05,4)] with confidence level of 95%. k=2.776

Thus, the confidence interval limit is:

(5.5)

Where: , is the standard deviation difference between TS and TLS

-47-

k: constant

: +, for the upper limit and -, for the lower limit intervals

Table 5.11: Confidence interval limit for the difference between TPS and TLS

Confidence interval limit and errors obtained from differences TPS and TLS

Abs[

]

Point N E Z N E H

BW10 0.034 0.027 0.028 0.002 0.001 0.001

BW11 0.015 0.012 0.034 0.002 0.001 0.007

BW12 0.017 0.015 0.005 0.002 0.002 0.001

BW13 0.009 0.017 0.017 0.001 0.002 0.001

BW15 0.007 0.012 0.010 0.001 0.001 0.001

P101 0.012 0.019 0.014 0.001 0.002 0.002

P102 0.131 0.019 0.017 0.012 0.002 0.001

P103 0.017 0.011 0.025 0.002 0.001 0.003

P104 0.017 0.037 0.016 0.001 0.004 0.002

P111 0.017 0.021 0.011 0.002 0.002 0.001

P112 0.016 0.029 0.010 0.002 0.002 0.001

P113 0.012 0.010 0.022 0.001 0.001 0.001

P114 0.016 0.014 0.010 0.001 0.001 0.001

RT1 0.010 0.017 0.015 0.001 0.001 0.001

RT2 0.018 0.019 0.009 0.001 0.003 0.000

RT3 0.016 0.011 0.009 0.003 0.001 0.001

RT4 0.024 0.018 0.032 0.001 0.001 0.002

WM1 0.020 0.024 0.018 0.002 0.002 0.003

WM2 0.013 0.019 0.012 0.001 0.002 0.002

WM3 0.024 0.044 0.019 0.001 0.003 0.003

WM4 0.017 0.013 0.015 0.001 0.001 0.001

As the result shows in Table 5.11, all values calculated from the coordinate difference between

total station and laser scaner lied within the allowable confidence interval limit.

-48-

5.6 Comparison of Time Expenditure

In order to compare the cost (time expenditure) of the methods applied, effective time has been

recorded throughout the measurements. Effective time refers to the time needed to measure the

required tasks without considering the delayed time due to some problems. The specified time is

specific to this measurement because it depends on the operator engaged. For the convenient of

comparison, time expenditure was classified in to time needed for total station versus GPS, and

total station versus laser scanner. The required time does not include the time for transportation

of instruments from store to the field and vice versa, and delayed time due to some problems

such as: battery problem, incorrect reading, etc.

5.6.1 Total Station versus GPS

As it is described in chapter one, the reference network consists of 14 control points, which were

measured from 7 stations towards all 14 points. This was done using two faces with two rounds

of measurements. The overall tasks were classified as field work and office lab work. But, here

the time consumed was recorded and compared only for the field measurement between TPS and

detail RTK on the reference network. Time allocated for every step of the measurement is

presented in Table 5.12. Time needed to setup the tripod of the instrument (total station) on one

station was recorded and then multiplied by the number of instrument stations to determine the

time expended on all instrument setups. In this project, the reference network has consisted of

seven instrument stations. The time expended for one setup of a tripod on one target is multiplied

by14 to calculate the expended time on tripod setup, since 14 is the number of control points in

the reference network. The same is true for the rest of measurement steps to determine the

expended time for total station measurement. Thus, the required total time with the total station

was 168 min (2 hours and 28 min). Time expended for GPS RTK was recorded as time required

for the reference base and for the rover. For the reference station, time was calculated as: time

required for tripod setup plus to center it which was 8 min. For the rover measurement, time has

been recorded as: time needed to center the rover plus time to record and to change to the next

station and then multiplied by the number of control points (13), totally it was 86min (1 hour and

26 min).

Finally, the overall expended time on the reference network using total station and GPS RTK

measurement has been recorded and compared. The time needed for the total station

measurement was 2 hours and 28 min and that of GPS RTK measurement was 1 hour and 26

min. When comparing required time of the two methods, total station was consumed more time

than RTK.

-49-

Table 5.12: Time expenditure for TS and RTK measurements for the reference network

Total station RTK

Time expenditure in (min)

Measurement steps Instrument Prism Reference Rover

Tripod Setup 4 4 5 ---

Centering 5 3 3 2

Aiming 3 --- --- ---

Recording 2 --- --- 2

Changing station --- 4 --- 2

Sum 14 11*14 8 6*13

168 86

5.6.2 Laser scaning and total station

In this sub section time comparison is presented in Table 5.13 between total station and laser

scaning measurement. In this measurement, both methods have shown somewhat comparable

time consumption. Time required for the scaning was recorded as: time needed for putting black-

white targets plus time for tripod setup plus time for aiming and recording, which was totally 32

min. Using the total station, required time was recorded as: time needed for tripod setup on

instrument station and back-sight point plus time for centering, aiming and recording. There were

21 points on the façade including the black-white targets. The required total time using total

station was 38 min. When comparing time consumed for those two methods, there was only 6

minute difference to accomplish each task. So, it was possible to measure with the scanner faster

than that of total station.

Table 5.13: Time expenditure for TS and laser scaner on the façade

Field

work

Façade measurement with total station Façade measurement with

laser scanner

Time expenditure in [min]

Tasks Instrument Targets/ prism Instrument Targets

Putting black-white

paper on façade

--- --- --- 20

Tripod Setup 4 4 5 ---

Centering 5 3 --- ---

Aiming 12 --- 2 ---

Recording 10 --- 5 ---

Changing station --- --- --- ---

Sum 31 7 12 20

38 32

-50-

6 CONCLUSION AND RECOMMENDATION

6.1 Conclusion

Today total station (TS), GPS and terrestrial laser scaner (TLS) are used for many tasks within

different applications, for example, geodesy, engineering, architectural and mining surveys and

documentation of cultural heritage with different accuracy level depending on the needed

requirements.

The purpose of this thesis work was to evaluate and compare accuracy, precision and time

expenditure of three surveying methods (TS, GPS and TLS). The comparison was made between

TS versus GPS RTK on the reference network and TS versus TLS on the façade of L building,

KTH campus, Stockholm, Sweden. To accomplish the objectives of the thesis, three major tasks

have been performed. 1. A network of 14 control points was established with high precision (l

mm) with total station and served as a reference or established value. 2. On the same network,

RTK method was performed to compare the result with that of total station. 3. Finally, by scaning

the targets on the façade of the L building and measuring the same target points with the total

station, comparison has been made between the extracted coordinates of the façade and the

coordinates measured by the total station.

In every task of the measurement, time expended was recorded and compared (see Table 5.13 and

Table 5.13) separately for TS versus GPS and TS versus TLS methods respectively. Then, in order

to evaluate the precision and accuracy of the RTK on the reference network and TLS and TS on

the façade measurements were taken five times.

Based on the results obtained, precision of the reference network determined with 1 mm standard

deviation both for horizontal and vertical coordinates for all points. This result has been achieved

because of the round measurements and two face measurements with the total station. On the

same control points of the network, RTK method was performed and according to the result

obtained, the standard deviations are less than 8 mm in horizontal and they reach 1.5 cm in

vertical coordinate, which indicates that the repeated measurements were quite close to each

other. The accuracy of the RTK measurements on the network, which is expressed by RMS, are

less than 9 mm in horizontal and they reach 2.2 cm in vertical coordinates.

Precision of the TS measurement on the façade of L building has been determined with

maximum standard deviation of 8 mm (point BW10) in horizontal and 4 mm (point WM3) in

vertical coordinates. On the same points of the façade, coordinates extracted from the TLS

measurement has been determined with maximum standard deviation of 1.6 cm (point WM3) and

1.2 cm (point BW11) in horizontal and vertical coordinates respectively. But the remaining

points were below this level. Then the accuracy of the TLS measurements was determined with

maximum RMS of 4 mm (point WM3) in horizontal and 7 mm (BW11) in vertical coordinates.

-51-

Finally, the time expenditure summarized as more time (82 min) was consumed for TS

measurement in the former method, but almost similar time (38 min for TS and 32 min for TLS)

was consumed in the latter case.

In order to evaluate the quality of the measurement, absolute value of each coordinate difference

between each method should not be exceed , which limits the errors not to be beyond certain

limit by multiplying their sigma differences with constant k (2.776). Based on this quality control

measure, more than 95% of the total result has achieved the requirement. This can be interpreted

as values which lied within the allowable limit (interval limit), considered as accepted values.

But values out of the interval limit considered as risk values, which might contain gross errors.

There was one point which was out of the interval limit and was rejected.

Therefore, it can be concluded that there were no gross errors in the measurement; because the

measurements were made precisely and accurately. For instance: when measuring using total

station, two face measurements was taken to eliminate some errors such as collimation axis

errors, tilting axis errors, etc. When using GPS RTK method, small tripod was used to erect the

rover vertical. Initially, I expected to achieve accuracy in mm level. But, due to some errors (like

centering error, instrumental error, satellite signal obstruction), some results have been deviated

in to cm level.

6.2 Recommendations

The obtained results from this thesis will hopefully improve the knowledge about accuracy,

precision and time consumption of the three methods used (TS, GPS and TLS). One can

differentiate which instrument should be used for which specific application depending on the

presented results. For further improvement of accuracy, the following recommendations are

forwarded:

Total station (Leica 1201) should be calibrated at some regular intervals. Since there was

problem in the level bubbles; one on the tribrach and the other on the total station

couldn’t be leveled at the same time. So, once calibrated the instrument, it will improve

the level of accuracy.

It can be achieved better accuracy by calibrating those instruments before the

measurement campaign.

Applications which require high precision so as to serve as reference value, such as

control point establishments, I recommend to use total station instead of GPS.

It was very difficult to manage the field measurement alone, specially establishing the

reference network has been a big problem. There will be a possibility of occurring gross

errors and therefore, I recommend working in group.

-52-

REFERENCES

Ahmed, E.M. (2012). Performance Analysis of the RTK Technique in an Urban Environment,

Australian Surveyor, 45:1, 47-54.

Angelo, J.B. (2004), Integration of Laser Scaning and Close Range Photogrammetry the last

decade and beyond, Ottawa, Canada.

ANZLIC (Australia and New Zealand Land Information Council), 2006. Glossary: spatial

information related terms. Retrieved on March 23, 2013 from

http://www.anzlic.org.au/glossary_terms.html.

Borgelt S C., Harrison J D., Harrison, K. A. Sudduth, and S. J. Birrell (1996). Evaluation of GPS

for Applications in Precision Agriculture; Appl. Eng. Agric. 12(6) 633–638.

Clark R.L. and Lee R. (1998). Development of Topographical Maps for Precision Farming with

Kinematic GPS; Trans. ASAE. Vol. 41(4) 909–916.

Csanyi, N. and Toth, C. (2007). Improvement of LiDAR Data Accuracy Using LiDAR Specific

Ground Targets. Ohio State University, Department of Civil and Environmental Engineering and

Geodetic Science, Columbus, USA.

Ehsani, M. R., Upadhyaya, S. K. and Mattson, M. L. (2004). Seed Location Mapping Using RTK

GPS; Trans. ASAE. Vol. 47(3): 909-914.

El-Hakim, S.F., and Beraldin, J.A. (2003). Critical Factors and Configurations for Practical

Image-Based 3D Modeling. In: 6th Conference on 3D Measurement Techniques. Zurich,

Switzerland. Sept. 22-25, Vol. II, pp. 159-167.

Fan, H. (1997). Theory of Errors and Least Squares Adjustment, KTH, Stockholm, Sweden.

Fazlay, F.S. (2003). LIDAR Image Processing Creates Useful Imagery, ArcUser Magazine,

University of Mississippi.

Forstner, W. (1994). Diagnostics and Performance Evaluation in Computer Vision. In:

NSF/ARPA Workshop: Performance versus Methodology in Computer Vision.

Fraser, C.S. (1987). Limiting Error Propagation in Network Design. Department of

Photogrammetric Engineering and Remote Sensing, University of Melbourne, Victoria,

Australia.

Fregonese, L., Pael A., van Schadewijk A. (2007). Structural Monitoring of a large Dam by

Terrestrial Laser Scaning, University of Milan, Italy.

-53-

Kostov, G. (2011). Using of both Fast Static and RTK Modes for GNSS Determinations to

Obtain Required high Accuracy and Productivity, According to the Current Possibilities of the

IT. Marrakech, Morocco.

Habib, A. and Schenk, T. (1999). A New Approach for Matching Surfaces from Laser Scaners

and Optical Sensors, Ohio State University, Department of Civil and Environmental Engineering

and Geodetic Science, Columbus, USA.

High Definition Scaner (HDS) Leica Geo-system, retrieved on March 20, 2013 from

http://hds.leicageosystems.com/en/5940.htm.

Charles K.T. and Jie S. (2009). Topographic Laser Ranging and Scanning Principles and

processing, retrieved on 20 March, 2013 from

http://www.asprs.org/a/publications/pers/2009journal/october/review.pdf.

Jonsson K.O., Andersson A., Jacobsson S.O., Vandevoorde S., Lambert D.M., Fowler C.J.

(2003), SWEPOS Network-RTK Services, status, applications and experiences. Presented at ION

GPS/GNSS 2003, 9-12 September, 2003, Portland, Oregon, U.S.A.

Lin, L.S. (2004). Application of GPS RTK and total station systems on dynamic monitoring land

use. Proceedings of the ISPRS Congress Istanbul, Turkey 2004.

Leica 1200+ total station manual.

Leica GPS 500 user manual.

Quintero, M.S., Alonzo A. and Scot R. (2008). Theory and Practice on Terrestrial Laser Scaning

Training material based on practical applications, version 4.

Pflipsen, B. (2006). Volume Computation: A Comparison of Total Station versus Laser Scanner.

Master thesis, Department of Technology and Built Environment, University of Gävle, Sweden.

Roe, D. (2008). Quality control. In P. Lavrakas (Ed.), Encyclopedia of survey research methods.

(pp. 650-652). Thousand Oaks, CA: SAGE Publications, Inc. doi:

http://dx.doi.org/10.4135/9781412963947.n422.

Satalich, J. and Ricketson, R. (1998). Field test of Trimble 4000 Real-Time Kinematic GPS

Survey System; Journal of Surveying Engineering. Vol. 124(1) 40–48.

Sjöberg, L.E. (2012). Lecture notes on the course AH2923 Global Navigation Satellite System

(GNSS), Division of Geodesy, KTH, Stockholm, Sweden.

Surveying, Britannica online, retrieved on 09 February, 2013, from

http://www.britannica.com/EBchecked/topic/575433/surveying.

-54-

Total Station. Wikipedia, the free encyclopedia. Retrieved on March 18, 2013 from

http://en.wikipedia.org/wiki/Total_station.

-55-

Reports in Geodesy and Geographic Information Technology

The TRITA-GIT Series - ISSN 1653-5227

14-001 Muhammad Zaheer. Kinematic orbit determination of low Earth orbiting satellites, using satellite-to-satellite tracking data. Master of Science thesis in Geodesy No. 3130. Supervisor: Milan Horemuž. March 2014.

14-002 Razmik Agampatian. Using GIS to measure walkability: A Case study in New York City. Master of Science thesis in Geoinformatics. Supervisor: Hans Hauska. April 2014.

14-003 Solomon Dargie Chekole. Surveying with GPS, total station and terresterial laser scaner: a comparative study. Master of Science thesis in Geodesy No. 3131. Supervisor: Milan Horemuž and Erick Asenjo. May 2014.

TRITA-GIT EX 14-003

ISSN 1653-5227

ISRN KTH/GIT/EX--14/003-SE